Digital Image Stabilization Method and Imaging Device Using the Same

1. A digital image stabilization method comprising: taking N frames of a target image with an exposure of T; dividing each of the N frames into predetermined blocks; selecting a reference frame from among the N frames by calculating the amount of motion between each frame and the other frames and selecting the one with minimal amount of motion; estimating a motion vector of each of the remaining frames based on a reference region of the reference frame, wherein the motion vector is a minimum among a sum of absolute differences (SAD) values each of which is obtained by comparing the reference region with a corresponding block in each of the remaining frames; and generating a result frame by combining a weighted frame, which is obtained by mapping a weight calculated based on the estimated motion vector to each block, and a color component frame.

2. The digital image stabilization method of claim 1 , wherein the dividing each frame into the blocks comprises dividing each frame into the blocks in multiple levels.

3. The digital image stabilization method of claim 1 , wherein the selecting the reference frame comprises: calculating an amount of motion (AOM) value indicating how much a block corresponding to the reference region in each of N−1 frames shifts from the reference region in each of the N frames; and selecting a frame giving a minimum AOM value as the reference frame.

4. The digital image stabilization method of claim 3 , wherein the selecting the reference frame further comprises checking an appropriateness of the reference frame using an R-square measure.

5. The digital image stabilization method of claim 1 , wherein the generating the result frame comprises: excluding a motion vector having a SAD value greater than a predetermined threshold value with respect each block in the reference frame; and generating a similarity weight based on SAD values of the remaining motion vectors of the reference frame.

6. The digital image stabilization method of claim 5 , wherein the generating the result frame further comprises: obtaining the weight for each block by multiplying the similarity weight by a smoothing factor after generating the similarity weight; generating the weighted frame by mapping weights to all blocks, respectively, of the reference frame; and generating the result frame by multiplying the weighted frame by the color component frame.

7. The digital image stabilization method of claim 1 , further comprising normalizing the result frame using the weighted frame.

8. An imaging device comprising: an image signal processor configured to receive and process raw data of each of N frames taken for an exposure time of T and to output RGB format data and Y data; a Y module configured to process the Y data of each of the N frames; a motion estimator configured to: divide each of the frames into predetermined blocks; select one of the N frames as a reference frame based on the Y data; and estimate a motion vector of each of the remaining frames based on a reference region of the reference frame as a minimum among a sum of absolute differences (SAD) values each of which is obtained by comparing the reference region with a corresponding block in each of the remaining frames; an RGB module configured to process and store the RGB format data of each of the N frames; and a result frame generator configured to generate a result frame by multiplying a weighted frame generated based on the SAD values by the RGB format data.

9. The imaging device of claim 8 , wherein the motion estimator compares amount of motion (AOM) values obtained for the respective N frames, and selects a frame giving a minimum AOM value among the N frames as the reference frame.

10. The imaging device of claim 8 , wherein the motion estimator comprises: a block generator configured to divide each of the frames into the predetermined blocks; a frame motion detector configured to compare the reference region with a corresponding block in each of the remaining frames; and a reference frame selector configured to select as the reference frame a frame comprising the reference region giving a minimum amount of motion (AOM) value corresponding to a total amount of motion of the corresponding block from the reference region with respect to the motion vector.

11. The imaging device of claim 10 , wherein the reference frame selector checks an appropriateness of the reference frame using an R-square measure to select the reference frame.

12. The imaging device of claim 10 , wherein the result frame generator generates the result frame by multiplying a weight based on the SAD values of motion vectors of the reference frame, by the RGB format data.

13. The imaging device of claim 12 , wherein the weight is obtained by a smoothing factor by a similarity weight generated based on the motion vectors except for motion vectors having a SAD value greater than a predetermined threshold value.

14. The imaging device of claim 13 , wherein the weighted frame is a result of mapping the weight to each of the blocks in the reference frame.